Proton therapy for adult craniopharyngioma: Experience of a single institution in 91 consecutive patients

Abstract

Background

Craniopharyngioma (CP) in adults is a rare benign tumor associated with many morbidities, with limited contemporary studies to define treatment, and follow-up guidelines.

Methods

A single-center retrospective study was conducted on patients aged ≥ 18 years from 2006–2018 with CP and who were treated with proton therapy (PT). Late toxicity was defined as a minimum of 18 months from diagnosis. Overall survival (OS), local recurrence-free survival (LRFS), and toxicity were characterized using Kaplan–Meier and Cox regression analyses.

Results

Ninety-one patients met the criteria, with a median age of 37 years (range 18–82 years). PT was conducted after tumor resection in 88 patients (97%), in 64 patients (70.3%) as an adjuvant strategy and in 27 (29.7%) after recurrent disease. Three patients received exclusive PT. A median MRI follow-up of 39 months revealed 35.2% complete response, 49.5% partial response, and 9.9% stable disease. Five patients developed local recurrence (LR). The pattern of failure study showed that these five LR were within the GTV volume. The 5-year LRFS was 92.0% [CI 95% 84.90–99.60]. All the patients were alive at the end of the follow-up. Patients requiring treatment adaptation during PT tend to have a higher risk of LR (P = .084). Endocrinopathy was the most frequent grade ≥ 2 late toxicity. Among patients who were symptom-free before the start of treatment, none developed hearing toxicity but four (9.8%) developed visual disorders and 10 (11.3%) symptomatic memory impairment. Patients with large tumors had a higher risk of developing symptomatic memory impairment (P = .029).

Conclusion

Adults with CP treated with PT have favorable survival outcomes, with acceptable late toxicity. Prospective quality-of-life and neurocognitive studies are needed to define late adverse effects better.

Craniopharyngioma (CP) is a rare, slow-growing World Health Organization (WHO) grade I neoplasm. It represents between 2 and 5% of all primary intracranial tumors¹ and has a typical bimodal incidence (first peak occurring from 5 to 15 years old and second peak from 45 to 60 years old).² For several decades, radical surgical resection has been considered the gold treatment standard³ with the goal of reducing the risk of neoplastic recurrence. Nevertheless, primary gross total resection (GTR) holds high mortality and morbidity risks and is often limited by the tight tumor adherences to vital hypothalamic nuclei, whose damage can lead to severe hypothalamic injuries.⁴ Therefore, current surgical planning is aimed to provide a maximal safe resection that spares hypothalamic morbidity and radiation therapy (RT) has emerged as an adjuvant strategy after safe sub-total resection (STR). Patients treated with safe STR followed by RT have similar survival outcomes to patients treated by gross tumor resection.^5,6 Moreover, as reported by a single institution study, hypothalamus-sparing surgical strategy combined with adjuvant RT decreased the rate of serious long-term obesity in childhood patients without increasing their risk for local relapses.⁷ Similarly, Schoenfeld et al. reported that the combination of STR and RT demonstrated a significantly lower risk of developing diabetes insipidus in an adult and child population of 122 patients. Nevertheless, STR with RT may be associated with a high risk of cognitive impairment, especially for younger patients which have a long life expectancy.⁸ While the good neurocognitive outcome has been described after 3D conformal RT,⁸ proton therapy (PT) offers a higher dose gradient and may be a suitable approach to further reduce the risk of radiation-induced toxicity in these patients, particularly in cochlea and brain parenchyma.^9,10 Indeed, in a uniform medium, monoenergetic protons travel a well-defined distance, losing energy at an increasing rate before coming to a halt. This forms the characteristic Bragg peak. Distal penumbra is limited and well adapted to the treatment of midline brain tumors. The main objective of the present study was to report the outcomes and proton-related toxicity of a large cohort of consecutive adult patients treated with PT for craniopharyngioma. Secondarily, this study aimed to analyze the factors that were conducted to PT failure and PT toxicity risk factors.

Methods

Patient Population

We retrospectively reviewed the medical records of all the 91 adult CP patients who received PT at Institut Curie between December 11, 2006, and December 28, 2018. Patients were enrolled on an Institutional Review Board-approved outcomes tracking protocol (IRB number: Adult CP—DATA220077). Patients below 18 years old at the time of PT were excluded. No patient had received prior radiotherapy. All the patients had pathologic confirmation of CP. The date of diagnosis was the date of the first confirmatory imaging study identifying the CP. The patients were included if they had received PT after gross total or sub-total resection, or if they had received radical PT (in case of non-operability) and if they had at least six months of follow-up after the end of proton therapy.

Surgical Management

The extent of surgery was defined as gross total resection (GTR), sub-total resection (STR), or biopsy. GTR was defined as complete tumor removal as reported by the operating neurosurgeon and confirmed by neuroradiological examinations (MRI). Surgical tumor debulking that did not result in GTR was defined as STR. Biopsy was defined as the surgical removal of a tumor sample for pathology analysis, with no further attempt at tumor resection. Cyst aspiration or marsupialization was also performed for patients with giant cysts before PT.

Proton Therapy

All patients underwent simulation for radiation treatment planning purposes, which consisted of the fabrication of a customized thermoplastic mask for immobilization in the supine position, followed by CT imaging with or without contrast enhancement. All patients treated before 2016 also had surgical implantation of four to five fiducial gold markers in the outer table of the skull in order to help the alignment. The target volume and normal tissue structures were delineated on each axial CT slice, supplemented with fused diagnostic MRI. Target volumes were generated as follows: (i) gross tumor volume (GTV) incorporated macroscopic tumor tissue whether cystic or solid or calcification, (ii) clinical target volume (CTV) was a 5-mm isotropic expansion to the GTV and was limited to 2–3 mm when CTV extended into normal-appearing brain parenchyma, (iii) planning target volume (PTV) was generated by adding a 2-mm isotropic expansion to CTV. Dosimetric plans were performed using the Isogray^® (Dosisoft, Cachant, France) proton treatment planning system. Proton therapy was prescribed to the PTV at a dose of 54 GyRBE over 30 fractions (1.8 GyRBE per fraction) for 57 patients, and 52.2 GyRBE over 29 fractions for 34 patients, in order to spare the optic chiasm. All treatment plans utilized double scattering proton therapy (DSPT). Proton plans generally consisted of five or six (87.8%) fields including right and left lateral fields, right and left superior oblique fields and a superior anterior oblique field (or, infrequently, a posterior-anterior field requiring seated positioning if in the fixed beam room). Beam angles were selected to avoid proton fields with an end-of-range in the brainstem and to reduce exposure of the temporal lobes. Individualized brass apertures and compensators were generated for each patient. Treatment planning goals included 100% of the CTV volume covered by 95% of the prescription dose. The following organ-at-risk (OAR) dose limits were met for all patients as recommended¹¹: middle brainstem maximum dose of 0.03 cc < 54 Gy RBE; optic nerve maximum dose of 0.03 cc < 55 Gy RBE; and optic chiasm maximum dose of 0.03 cc < 52.2 Gy RBE. Protection of the chiasma in DS delivery was performed for the last fraction with an adapted collimator. Daily orthogonal radiographic imaging based on bony anatomy was used for daily image guidance, with a six degree of freedom registration. For cystic tumors, MR images especially the T1 and T2-weighted sequences, and the 3D fat-suppressed T1-weighted MR, or/and contrast enhanced CT were performed during radiotherapy every two weeks, to assess for tumor/cyst changes and to decide the need to replan.

Follow-Up

After the end of the PT, follow-up visits were conducted every three months for the first two years, and every 12 months thereafter or until death. Multidisciplinary follow-up involved neurosurgeons, radiation oncologists, endocrinologists and ophthalmologists. Each visit included a clinical physical examination, a biological assessment of pituitary function (including the exploration of gonadotropin deficiency, thyrotropin deficiency, corticotropin deficiency, growth hormone deficiency, and diabetes Insipidus), a brain MRI, an audiogram and ophthalmological assessment (in particular visual field evaluation) as recommended.¹² Local recurrence (LR) was detected by clinical examination and imaging. Recurrence or disease progression was defined as enlargement of the cystic and/or solid components of the tumor (confirmed by MRI) at any time after surgery alone or 1 year after PT. For cystic forms, progression had to be confirmed as continuous with at least two successive MRI. The Common Terminology Criteria for Adverse Events, Version 5.0 (CTCAE v5.0), formerly called the Common Toxicity Criteria (CTC or NCI-CTC), were used to assess the late toxicities of RT at each patient visit. As usual, toxicities were classified as late if they occurred at least 90 days after the end of PT. It should be noted that formal neuro-cognitive assessments were not performed. Memory impairment and neuro-cognitive dysfunction were mainly reported by family members and, more rarely, by the patient himself. These toxicities were also qualitatively estimated by the radiation oncologist and the neurosurgeon during follow-up visits when psychomotor slowing was observed.

Pattern of Failure Study

For this section, we used the same approach already applied in three previously published studies.¹³ For all patients with LR after PT, the recurrent tumor volume (Vrecur) was identified on MRI obtained at the time of recurrence diagnosis. The exact site and extent of each tumor were then compared visually to the pretreatment planning CT datasets. The recurrences were categorized as occurring inside or outside the previously irradiated targets: the Vrecur was deemed “in-field”, if the majority of Vrecur was within the GTV; “marginal,” if ≤ 50% of Vrecur was within the GTV; or “outside,” if less than 20% of the Vrecur was inside the GTV.

Statistical Analysis

Baseline characteristics were summarized as numbers and percentages for qualitative data, and as medians with the inter-quartile ranges for continuous variables. Follow-up was calculated from the date of the end of the PT to the date of the last follow-up. The median follow-up was estimated by the Kaplan–Meier method. Overall survival (OS) was defined as the time between the date of the end of the PT and the date of death. Patients still alive were censored at their date of the last follow-up. Local recurrence-free survival (LRFS) was calculated from the date of the end of the PT until the date of LR. In the absence of local recurrence, patients were censored at the date of their last follow-up. Survival curves were computed using the Kaplan–Meier method and compared using the log-rank test. Kaplan–Meier method was also used to perform univariate analyses on toxicity. For each toxicity (visual, cognitive, memory or one of the components of the hormonal axis) a delay to toxicity was calculated from the end of PT to the date of the first observed grade 2 toxicity. The delays served to compute curves to toxicities occurrences. Age, number of surgeries, GTV size and tumor extension were evaluated by univariate analyses as prognostic factors of the toxicities using the log-rank test. A threshold search was carried out to divide the GTV size quantitative data into two classes of qualitative variables: tumor ≤ 3.7 cc or > 3.7 cc. All tests were realized at the bilateral threshold of 5%. Analyses were carried out using R software version 4.1.1.¹⁴

Results

Patients’ Characteristics

Ninety-one adult patients with craniopharyngioma were treated with PT and were included in this study (Table 1). The median age at diagnosis was 34 years old (IQR, 23–49 years). The median age at the time of PT was 37 years old (IQR, 26–51 years). The most common presenting symptoms were endocrinopathies (affecting 82.4% of patients) and vision changes (affecting 54.9% of patients) (Supplementary Data 1). Most patients had the adamantinomatous subtype, while 13.2 % had the squamous-papillary subtype. It should be noted that none of the cases included in this series was malignant. The “suprasellar” extension was defined using the Prieto et al.’s classification.¹⁵ Sixty-four patients (70.3%) were treated at initial presentation and 27 patients (29.7%) were treated for recurrent disease. Among the 64 patients treated at initial presentation, three patients received definitive PT and 61 received PT after sub-total resection (STR). Among the 27 patients treated for recurrent disease, 11 patients received PT after GTR and 16 patients received PT after STR (Supplementary Data 2). The median time to recurrence from prior surgery for the 27 patients undergoing salvage treatment was 14.5 months (IQR, 10–24).

Proton Therapy Details

The median maximum dose prescribed to the GTV was 54 Gy (IQR, 52.7–54 Gy). The median overall treatment time was 42 days (IQR, 41–45 days). The median CTV and PTV volumes were 13.7 cc (IQR, 9.4–22.4) and 22.7 cc (IQR, 16.3–34.7), respectively (Table 2). Thirty-one patients underwent MR or/and CT evaluation of the target volume during PT and five of them required treatment adaption for cyst growth.

Disease Control and Survival Outcomes

With a median follow-up of 39 months (range, 7–147 months), the radiological response assessed on MRI was complete, partial or stable for 32 (35.2%), 45 (49.5%) and 9 (9.9%) patients, respectively (Figure 1). A total of five patients (5.5%) developed LR: two cystic forms and three mixed forms (with solid and cystic components). The three-years and five-years LRFS were 94.79% [95% CI 89.91–99.94] and 92.0% [CI 95% 84.90–99.60], respectively. Among the 11 patients who received PT after GTR, none recurred. Figure 2 shows that the five recurrences occurred within the irradiated GTV. All the patients were alive at the end of the follow-up. Among the five patients who required treatment adaptation, three had a LR. In univariate analysis, age at PT, sex, pathological subtype, tumor size, tumor extension, radiological presentation at the diagnosis (cystic or solid), and the time interval between the first and the last surgery and the onset of PT were not significantly associated with local control. Patients irradiated after more than one surgery (Supplementary Data 3) and who had treatment adaptation during the PT (Supplementary Data 4) seemed to have a higher risk of LR (P = .078 and P = .084, respectively). It should be noted that 4/5 LRs occurred in the subgroup of craniopharyngioma involving the third ventricle and/or the hypothalamus.

Toxicity

Fatigue and headaches were the most frequent acute toxicities observed in our cohort. Twenty-seven patients (29.7%) experienced grade 2 acute toxicity (Table 3, upper part). All the 91 patients experienced grade ≥ 2 late toxicity, in particular endocrinopathy (92.3%), including corticotropin, thyrotropin and diabetes insipidus (Table 3, upper part). However, we have shown that most patients included in this study already had endocrinopathy or visual impairment at the time of diagnosis (Table 2). To better define radiation-related toxicity, we investigated a subgroup of patients who had no symptoms before starting radiotherapy (grade CTCAE v5 0 or 1) and developed grade ≥ 2 late toxicity after radiotherapy (Table 3, low part). The most frequent and severe grade 2 toxicity in this subgroup were corticotropin and thyrotropin deficiencies which concerned 46.7% and 36.8% of patients, respectively. The median time to onset of these toxicities was 13 months (IQR: 3.75–20) and 19 months (IQR: 13.5–27.25), respectively. Among the 4 patients who were symptom-free before the start of treatment and developed late grade ≥ 2 visual disorder, one experienced 5 months after PT visual impairment evolving toward grade 4 visual loss. A dosimetric study was conducted and showed that dose constraints for visual pathways were respected. Among the 11 patients who received PT after GTR, none developed memory or cognitive disorder. Among the four patients who received PT after GTR and did not have visual dysfunction before PT, none developed visual impairment. Supplementary Data 5 summarizes the outcomes of patients who developed late grade ≥ 2 toxicity after PT. The number of patients with persistent toxicity at the last follow-up was very limited. None of the patients included in our series had hearing loss or stroke after PT.

For each grade ≥ 2 late toxicity, univariate analysis was performed to evaluate whether the age at PT, the size of the target volume, the extension (Prieto et al.’s classfication¹⁵) and the number of surgeries could be associated with the risk of late toxicity. None of these three factors were significantly associated with the risk of visual impairments, cognitive trouble or endocrinopathy. However, the risk of memory impairment was significantly associated with GTV size: patients with a tumor larger than 3.7 cc had a higher risk of developing late symptomatic memory impairment than patients with a tumor smaller than 3.7 cc (P = .029, Figure 3).

Discussion

The present study allowed us to assess the outcomes of a large cohort of adult patients irradiated with proton therapy for craniopharyngioma (CP). Despite the bimodal incidence of CP data on adult CP are scarce and limited compared to those of childhood patients in the literature.¹⁶ There are some differences between the adult and child populations with CP. The papillary subtype occurs almost exclusively in adults comprising about 14–50% of the tumors in this age group.^6,17 This histopathological subtype rarely presents calcifications, is usually well circumscribed, and compared to the adamantinomatous type, surrounding tumor infiltration is less frequent.¹⁸ Moreover, the included patients also had predominantly small tumors (median, IQR): 3 cm (1.6–7.15), which is in agreement with the recent meta-analysis by Lehrich et al. that showed that adult patients had a majority of tumors less than 3 cm in size compared to pediatric patients.¹⁷ In addition, vision changes and endocrinopathy were common signs at the time of diagnosis before proton therapy (PT) in our series. Previous reports had indeed shown that in adults and in children the most common presenting clinical symptoms were visual field deficits and signs of hypopituitarism.^19,20

In our knowledge, there are currently no evidence-based guidelines or a clear consensus for the best treatment of primary or recurrent CP in adults. Like in children a radical approach with complete tumor resection and a potential cure has to be balanced with a more conservative approach to avoid substantial treatment-associated long-term morbidity.²¹ In particular, gross total resection (GTR) may result in significant and devastating peri- and post-operative morbidity especially after resection of tumors invading the hypothalamus.²² A recent meta-analysis showed that the risk of recurrences after GTR and sub-total resection (STR) followed by radiation therapy (STR with RT) was not significantly different.⁶ In this study, the risk of developing recurrence was significant for GTR vs STR (odds ratio [OR]: 0.24, 95% CI 0.15–0.38) and STR + XRT vs STR (OR: 0.20, 95% CI 0.10–0.41). Comparison of the risk of recurrence after GTR vs STR + XRT did not reach significance (OR: 0.63, 95% CI 0.33–1.24, P = .18). Similarly, Zhang et al. reported the outcomes of 1218 patients with CP treated with RT and/or surgery (GTR or STR).²³ Overall survival/cause-specific death for patients that underwent RT, STR + RT, and GTR were not statistically different. In our study, the majority of the patients (84.6 %) received RT after STR. The 5-year local failure-free survival and the overall survival were excellent: 92% and 100%, respectively. This is consistent with a recent study, in which Rutenberg et al. reported the outcomes of 49 adult patients treated with RT for CP, 77% of whom received postoperative RT.²⁴ With a median clinical and radiographic follow-up of 4.2 (range, 0.4–21.6) years and 3.0 (range, 0–21.5) years, the 5- and 10-year local control rates were 100 and 94%, and the 5- and 10-year overall survival rates were 80 and 66%.

Moreover, it should be noted that the five patients with LRs in our series had at least two surgeries before the PT. These were ultimately patients with large tumors (> 7 cc) and, for 4/5 patients, tumors that involved the third ventricle and/or the hypothalamus. These tumors are indeed known to be difficult to access and often require two volumetric reduction surgeries (high and low approach), with nevertheless a high risk of LR.⁴ These patients should be closely monitored. Besides, all the five LRs observed in our study occurred within the GTV. Therefore, we hypothesize that the use of an isotropic margin of 5 mm to construct the CTV could be reduced, which would induce a reduction in the volume of healthy tissue irradiated and thus a decrease in the risk of toxicity.

Several studies have shown cyst growth during radiotherapy,^25,26 possibly related to a portion of the GTV that has not received the full prescribed dose.²⁷ Thus, especially in pediatric patients, radiation oncologists have developed a systematic approach to assessing the size of cysts during RT, by performing imaging, usually MRI, every 2 weeks during treatment.²⁸ If there is significant cyst growth, most teams opt to replan. In our studies we have observed that, despite this attitude, patients who had treatment adaption had more LRs than patients who did not have this adaption (Supplementary Data 4). These are therefore potentially fast-growing tumors that require even more frequent assessment of their size during PT. This is not obvious in PT because the machines are not all designed with cone beam CT on board, as is the case for most photon beam devices. An alternative would be to anticipate cyst growth at the time of initial treatment planning by adding a larger CTV margin. Some innovations in imaging and radiomics would be useful in predicting this cyst growth.²⁹ These patients, for whom the risk of cyst growth is high, could also be treated on the new MRI-guided radiotherapy devices, as the treatment can be adapted in real time.³⁰

In the present study, a large majority of patients had grade ≥ 2 toxicity at the last clinical follow-up. Nevertheless, it should be noted that most of them had visual changes and endocrinopathy before the start of RT. Among the patients who did not have visual impairment before RT, 9.8% developed visual disorders, which is in agreement with the 11% of visual disorders observed in Rotenberg et al. ’s study.³¹ Several studies have already reported the risk of late endocrinopathy after radiotherapy for CP, either in proton or in photon.³² Since the pituitary gland is totally or largely irradiated at the maximum prescribed dose (54 Gy), it is almost completely irrelevant to limit the risk of endocrinological toxicity. It is therefore recommended that biological tests be performed at regular intervals before and after radiotherapy in order to adapt treatments for these pituitary deficits.¹² Besides this, as Merchant et al. shown in 2008, PT may be also useful to spare cochlea and decrease the risk of hearing loss.³³ In our series, no patient presented late hearing loss.

Moreover, in our series, among the patients who did not have cognitive impairment (90/91) and memory impairment (89/91) before the onset of the PT, 4.5% and 11.3% developed grade 2 late cognitive and memory disorders, respectively. Actually, one of the main clinical benefits of PT in CP is expected to improve neuro-cognitive outcomes.¹² In 2017, Toussaint et al. compared cognitive test results from two prospective trials that included children with CP treated with proton (NCT01419067) or photon (NCT00187226).³⁴ When corrected for the distribution of radiation dose in normal brain, those treated with double scattering proton therapy (DSPT) had no change in academic achievement scores (reading and math) compared with patients treated with photon therapy, who showed a significant decline. Several dosimetric studies were also conducted to compare patients with brain tumors, in particular CP, the dose delivered with proton or photon in the temporal lobe.^33,35–37 Merchant et al. reported for 10 child patients with CP a significant decrease of the low doses in the temporal lobes when using DSPT compared to photons.³³ Moreover, the recent Toussaint et al.’s study reported and compared the dose received by 30 brain substructures associated with cognition with volumetric modulated arc therapy (VMAT), double scattering proton therapy (DSPT) or pencil beam scanning (PBS) PT.³⁷ The exposed volumes of the temporal lobes and their substructures were consistently reduced with PBS compared to DSPT and VMAT: the left hippocampus V10Gy was from 100% (VMAT) or 41% (DSPT) to 5% with PBS (P = .002). The reduced doses to the temporal lobes achieved with PBS translated into improved predicted memory outcomes. It should also be noted that the relation between the temporal lobe, more specifically hippocampal dosimetry, and memory decline is still under debate.^11,38,39 Hippocampal dose constraints (D40% < 7.3 Gy) were based on Gondi et al. ’s study in which authors shown that conformal avoidance of the hippocampus during whole brain RT was associated with preservation of memory and quality of life as compared with historical series without hippocampal preservation.^39,40 However, several authors have recently demonstrated that cognitive and memory changes should be also due to the irradiation of others structures such as cerebral white matter, subventricular zones, or cerebral cortex.⁴¹ Consequently, only a prospective study evaluating in an objective way the association between the dose delivered to all these brain regions and the symptoms of cognitive decline and memory, assessed using neuro-cognitive scores, would allow to providing an appropriate response to this question. Finally, it is interesting to note that in our series, patients with large tumors had significantly more long-term memory impairment than patients with smaller tumors. One explanation could be that at the level of the lateral wall of the 3rd ventricle, the memory pathways are located about 3 mm from the ependyma. If the residue was located at the level of the wall, there was therefore an increased risk of post-PT memory disorders, especially since the CP is most often infiltrating, invasive at this level, with tumor digitations within a reactive gliosis.

To limit tumor size before PT, and thus decrease this risk of late toxicity, medical debulking with BRAF/MEK inhibitors could be considered for papillary CP. Indeed, while adamantinomatous CP is driven by CTNNB1 mutations, papillary CP is associated with mutations in BRAFV600E with the subsequent upregulation of mitogen-activated protein kinase (MAPK). This latest finding offers the possibility of using BRAF/MEK inhibitors to reduce the tumor mass of papillary CP and thus the toxicity associated with PT in this context, more frequent in the adult population.^42,43 To our knowledge, there is currently no indication for medical treatment adjuvant to PT for CP. Identification of poor responders who especially require adaptation during PT and who carry a mutation should be an indication of prophylactic BRAF/MEK inhibitors.⁴⁴

The main limitations of our study were the retrospective recording of data, especially for late toxicity. Neuro-cognitive changes were not assessed using neuro-cognitive scores. Hypothalamic symptoms were also not explored (body weight, sleep-wakefulness cycle, emotional and psychiatric alterations). Therefore, a prospective study comparing GTR and STR followed by PT, including prospective recording of late toxicity with neurocognitive and psychiatric scores, would be useful to validate the results observed in our series. Another limitation is the relatively short follow-up with a median of 39 months (range, 7–147). Moreover, in the present study, at least 13.2% of patients had papillary CP, which is relatively low compared with the other studies previously cited that evaluated the outcome of adult patients with CP. The papillary subtype is overall less aggressive than adamantinomatous subtype and this could have consequences on the outcome of patients. Nevertheless, 13.2% of papillary subtypes in our cohort is equivalent to that of Rutenberg et al. who reported equivalent results.³¹ Furthermore, the pathologic subtype was not specified in 45% of the pathologic analyses in our cohort. Therefore, it is possible that this proportion was underestimated.

Finally, this study demonstrated in a large cohort of patients that PT is a highly effective modality for the treatment of CP in adults. Late endocrinopathy remains frequent, but visual changes, especially in patients who were not visually impaired before PT, were relatively rare. Cochlea and temporal lobes were spared and the incidence of hearing loss and severe memory/cognitive disorders were very limited. Thus, this study confirmed that STR followed by PT is a valid option for the treatment of CP.

Conflict of Interest statement

None.

Funding

None.

Authorship statement

Conception and design: AB, NS, CA, VC, LF. Collection and assembly of data: AB, NS, GB, LF. Data analysis and interpretation: AB, NS, SD, VC, LF. Manuscript writing: all authors. Final approval of manuscript: all authors. Author responsible for statistical analysis: Sylvain Dureau, Department of Statistics, Institut Curie, Saint-Cloud ([email protected]).

References